Children's selective trust in technological informants: A meta-analysis

Children increasingly encounter technology as a source of information, yet evidence of whether they trust technological informants more or less than human informants is mixed. The present meta-analysis synthesizes 20 studies (27 experiments; N = 8,772; ages 3-10 years) to examine children's selective trust in technological versus human informants and to identify theoretically motivated moderators of this trust. Across studies, there was no overall preference for technological or human informants. However, substantial heterogeneity emerged. Multivariate metaregression revealed four significant moderators: age, information domain, informant type, and experimental paradigm. Relative to 3-4-year-olds, older children showed greater trust in technological informants. Children trusted technology more for learning about novel words and objects than for learning about humans and biology-related information. Children's trust in repository-based tools (e.g., computers, internet search engines) compared with humans was greater than their trust in conversational agents (e.g., robots, voice assistants) compared with humans. Children in studies using single-informant paradigms favored humans more than those in studies using conflicting-source paradigms. Informant characteristics (e.g., prior accuracy, benevolence) and the type of trust measure (ask vs. endorse) did not reliably moderate effects. Together, these findings suggest that children's trust in technology changes with age and context, and it is best captured by an epistemic-calibration account, although social-communicative cues may still play a role in children's judgments. (PsycInfo Database Record (c) 2026 APA, all rights reserved).